J.Med. Chem. 1994,37, 415-420
416
Synthesis and Photodynamic Activities of Silicon 2,3-Naphthalocyanine Derivatives Nicole Brasseur, Tan-Loc Nguyen, Rejean Langlois, Rene Ouellet, Stephanie Marengo, Daniel Houde, and Johan E. van Lier' MRC Group in the Radiation Sciences, Faculty of Medicine, University of Sherbrooke, Sherbrooke, Qukbec, Canada J1H 5N4 Received September 28, 1996
Bis(tert-butyldimethylsi1oxy)-(7), bis(dimethylthexylsi1oxy)- (8), bis(tri-n-hexylsi1oxy)- (9), and bis(dimethyloctadecylsi1oxy)silicon 2,3-naphthalocyanines (10) were prepared via substitution of the bis(hydroxy) precursor with the corresponding chlorosilane ligands and characterized by spectroscopic and combustion analyses. They show strong absorption around 780 nm where tissues exhibit optimal transparency. Compounds 7-10 are capable of producing singlet oxygen. They are relatively photostable although less stable than the analogous phthalocyanine, i.e., the bis(dimethylthexylsi1oxy)silicon phthalocyanine (12). They were evaluated as potential photosensitizers for the photodynamic therapy (PDT) of cancer in uitro against V-79 cells and in vivo against the EMT-6 tumor in Balb/c mice. In uitro all four dyes showed limited phototoxicity combined with substantial dark toxicity. Surprisingly, in uiuo (iv, 0.1 pmol/kg, 24 h prior to the photoirradiation of the tumor with 780-nm light, 190 mW/cm2,400 J/cm2) all dyes induced tumor regression in a t least 50% of mice whereas compound 8 gave a complete tumor response in 80% of mice without apparent systemic toxicity a t doses as high as 10 pmoVkg. At 24 h postinjection, compound 8 showed a favorable tumor to muscle ratio of 7, assuring minimal damage to the healthy tissue surrounding the tumor during PDT. Our data confirm the potential of silicon naphthalocyanines as far-red-shifted photosensitizers for the PDT of cancer and indicate the importance of the selection of the two axial silicon ligands for optimal photodynamic efficacy. Introduction Clinical trials for the treatment of a variety of cancers with red light after systemic administration of a mixture of hematoporphyrin derivatives (Photofrin), known as photodynamic therapy (PDT), are in progress in medical centers throughout the world, and in Canada approval of this protocol has been secured for the treatment of bladder cancer. It is believed that PDT has the potential to evolve as a valid alternative or adjuvant therapy for a variety of solid neoplasms.' Over the past years, several alternative classes of dyes with physicochemical properties selected for improved efficiency of PDT have been proposed as new sensitizers for PDT.2 Among them, the phthalocyanines (Pc) and naphthalocyanines (Nc) have received increasing a t t e n t i ~ n .Compared ~ to Photofrin (e N lo3 M-l cm-l), these dyes offer high molar extinction coefficients (6 N 106 M-I cm-l) and red-shifted absorption maximum at 680 (Pc) and 780 nm (Nc) resulting from the benzene rings condensed to the periphery of the porphyrin like macrocycle. Like Photofrin, Nc and Pc are capable of generating singlet oxygen ('OZ),the reactive species generally believed to be responsible for the cytotoxiceffect. The use of 780-nm red light in the case of Nc allows for light penetration into tissues to twice the depth of that possible at 630 nm, Le., the wavelength currently used for porphyrin-mediated PDT.4 Convenient light sources are available in the far-red-near IR region of the spectrum, including the solid-state diode and tunable Ti-Sapphire lasers. The singlet oxygen yields obtained with Nc are comparable to those reported for Pc and porphyrins;5 however, compared to the analogous Pc, Nc are more prone to photobleaching.6 Phthalocyanines and naphthalocyanines form stable chelates with metal cations, and the photosensitizing properties of aluminum, zinc, and silicon naphthalocya*Abstract published in Advance ACS Abstracts, January 1, 1994.
0022-2623/94/1837-0415$04.50/0
nines have been documented.' We now report on the potential of four silicon naphthalocyanine derivatives (SiNc) as photosensitizers for PDT. The choice of silicon as the central metal ion was based on its tetravalence which allows for the substitution with two axial ligands, whereby variations in the nature of the latter provide a base for the evaluation of structure-activity relationships. Furthermore, addition of bulky substituents to these molecules should reduce their tendency to aggregate, favoring the monomeric and photoactive form of the dye. Four axial substituents with increasing lengths of aliphatic chains (C4-Cl8) were selected, and the syntheses of the following derivatives are reported: the bis(tert-butyldimethylsiloxy)- (71, bis(dimethylthexylsi1oxy)- (81, bis(tri-n-hexylsi1oxy)- (9), and bis(dimethyloctadecylsi1oxy)silicon 2,3naphthalocyanines (10) (Scheme 1). Dyes were formulated as Cremophor EL emulsions, and their photodynamic activity was evaluated in uitro against V-79 cells and in vivo against the EMT-6 tumor implanted intradermally in BALBIc mice. Pharmacokinetics of the most active dye 8 were studied in the same animal model. Results Chemistry. The synthesis of the naphthalocyanine (Scheme 1) and phthalocyanine (Scheme 2) derivatives was adapted from established procedures.8 The dihydroxysilicon naphthalocyanine and phthalocyanine were selected as precursor for the introduction of substituents of different hydrophobicity and bulkiness. Briefly, a,a,a',a'-tetrabromo-0-xylene(1) was reacted with fumaronitrile (2) at 75 "C,using sodium iodide as scavenger for bromine, to yield 2,3-dicyanonaphthalene (3). This product was condensed with NHs in refluxing methanol in the presence of sodium methoxide as catalyst to give the l,&diiminobenz[flisoindoline (4).8b Compound 4 was then heated with tetrachlorosilane, tri-n-butylamine, and tetrahydronaphthalene at 200 O C for 4 h. The resulting 0 1994 American Chemical Society
Brasseur et al.
416 Journal of Medicinal Chemistry, 1994, Vol. 37, No. 3
Scheme 1
I
Om20
2
1
3
NH
It
R
0.00
II
*
600
NH
5-10
4
6!io
700 750 800 WAVELENGTH (nm)
850
Figure 1. Absorption spectra of a 0.4pM solution of compound 9 in DMF (dotted line) and culture medium containing 1% serum in the presence (bold line) or absence (Tie line) of 0.04% Cremophor. Table 1. Rates of Photobleaching and Quantum Yields of Singlet Oxygen dye solvent rate of ~hotobleaching,~ mol 8-l 9 (IO& 7 DMF 4.1 X 106 0.18 8 DMF 3.6 X 106 0.33 9 DMF 2.4 X 106 0.32 PBS-H20 4.3 x 10-8 PBS-D20 1.8 x 106 10 DMF 3.6 X 106 0.33 12 DMF 8.0 x 10-8 4 First-order rate of decay extrapolated to t = 0 during exposure of various SiNc and S P c (10pM) to 20-400 J/cm2 of red light ( A > 600 nm). Quantum yields for the production of singlet oxygen, 9 (~OZ),were measured in oxygen-saturated DMF at 2 pM dye concentration. The luminescence intensity of '02at 1276 nm was quantified using 9,lO-diphenylanthracene 88 a scavenger.9 ~
Scheme 2
*
HO OH
11
12
dichlorosilicon naphthalocyanine (5) was removed from the reaction mixture by filtration and converted to the corresponding dihydroxy derivative 6 by treatment with sulfuric acid and subsequent reflux in concentrated ammonium hydroxide. The dihydroxysilicon naphthalocyanine 6 was reacted with tert-butyldimethylchlorosilanein 2,4,6-collidineand tri-n-butylamineat reflux for 15h to give 7 as a dark green product. Compounds 8-10 were prepared in a similar manner using the appropriate chlorosilane reactants. All four compounds were purified by aluminum oxide column chromatography and characterized by their spectral properties. For comparative studies, a phthalocyanine derivative containing the thexyl substituents was also synthesized. The condensation of 1,3-diiminoisoindoline with tetrachlorosilane in quinoline at 200 "C afforded the dichlorosilicon phthalocyanine,& which was converted to the corresponding dihydroxysilicon phthalocyanine 11 with aqueous sodium hydroxide in pyridine.M The compound 11 was reacted with dimethylthexylchlorosilane in 2,4,6collidine and tri-n-butylamine, as described for the analogous Nc derivative 8, to give 12. Figure 1shows the absorption spectra of compound 9 in dimethylformamide (DMF) and in culture medium containing 1%serum with and without 0.4% Cremophor,
which served as injection vehicle for the dye. The sharp absorption maximum around 775 nm suggests that the dye is in a monomeric state, even in aqueous solutions. In the absence of Cremophor, the position of the maximum wavelength was not significantly shifted although its intensity was diminished. No important differences in photostability were found among the four SiNc derivatives diluted in DMF and exposed to light doses up to 100 J/cm2 (Table 1). The bis(dimethylthexylsi1oxy)silicon phthalocyanine 12 was about 7 times more photostable than the corresponding naphthalocyanine 8. The photostability (up to 400 J/cm2) of compound 9 was also studied in phosphate-buffered saline (PBS)containingO.5 % Cremophor. When the same experiment was performed in PBS formulated in DzO instead of H20, the rate of photobleaching of 9 increased about 3-fold (Table 1). Quantum yields for the production of singlet oxygen were taken from Marengo et aLgand are listed in Table 1. No difference in hydrophobicity among the four SiNc derivatives was observed when partition in the system octanol/tris buffer (1:l) was used as a criterion; all dyes were completely recovered in the organic phase (data not shown). However, when the hydrophobicity was related to the migration distance on TLC (Rf)or the retention time on HPLC ( t ~ a) good , correlation was observed with the length of the alkyl chain of the axialsubstituents (Table 2). Biological Properties. The effect of SiNc derivatives and light on the survival of V-79cells is expressed as the extracellular dye concentration (pM) required to achieve
Silicon 2,3-Naphthalocyanine Derivatives
Journal of Medicinal Chemistry, 1994, Vol. 37, No. 3 417
Table 2. Light and Dark Toxicity toward V-79Cells and ChromatographicMobilities of Differently Substituted SiNc LDma
(-light),
LDma
RP
tRb
(HPLC), dye Substituents) uM uM (TLC) min . . - (axial 7 (tert-butyldimethylsiloxy) 36 12 0.16 9.2 8 (dimethylthexylsiloxy) 6.5 5.5 0.18 9.0 9 (tri-n-hexylsiloxy) 46 29 0.53 5.0 10 (dimethyloctadecylsiloxy) 4.2 1.8 0.34 5.1 a Extracellular drug concentration (LDw) required to kill G % o f V-79cells after a 1-hincubationperiod (-light) and after subsequent exposure to 2.4 J/cm2 red light (+ light). Dye mobility relative to the solvent front (Rf)upon silica gel TLC in petroleum ether/ chloroform (3:l)and retention time ( t R ) on normal-phase HPLC in petroleum ether/chloroform (191). (+ light),
*
100 1
I
7
8
9
1
0
SING DERIVATIVES
Figure2. EM"-6 tumor response in BALB/c mice (n L 8)treated with 780-nm laser light (190 mW/cm2,400 J/cm2) 24 h after iv administration of 0.1 pmol/kg SiNc derivatives 7-10. Shaded area, flat and necrotic tumor within a few days after phototreatment; dark area, absence of a palpable tumor within 2 weeks after phototreatment.
50% cell kill (LDm, Table 2). All four SiNc derivatives 7-10 exhibited substantial dark toxicity. Exposure to light further reduced the cell survival to varying degrees with the exception of compound 8 which induced similar toxicity with or without light. A correlation between increasing hydrophobicity of the dye and efficiencyto inactivate cells (LDm) is noted, with the exception of compound 9 (Table 2). The latter is the most hydrophobic dye of the series but also the least active in terms of potential to induce photosensitized cell killing. The concentrations of dimethylthexyl derivative 8 in plasma, tumor, skin, and muscle at different time intervals after iv administration (1pmol/kg) to tumor-bearing mice are presented in Table 3. Plasma clearance follows firstorder kinetics, with a half-time for elimination of around 7 h. The tumor accumulated significant amounts of photosensitizer, reaching a maximum at 1 2 h after administration, while dye elimination was slow. Dye uptake by the muscle was very low, providing excellent tumor to muscle ratios reaching a maximum of 10 at 12 hpostinjection. However, at this time point, plasmalevels are still high, suggesting that the best time for the photodynamic treatment of the tumor is between 24 and 72 h after dye administration. Tumor dye levels also exceeded those of the skin throughout the study (tumor to skin ratio: 1.5-1.8). High levels of dye were retained by the liver and the spleen even at 1week after injection (data not shown). The tumor response upon exposure to red light (780 nm, 190 mW/cm2, 400 J/cm2) 24 h after iv injection of compounds 7-10 is summarized in Figure 2. Control animals exposed to light alone showed no apparent tumor effect. Also, nonirradiated control tumors of animals injected with dye showed no regression. At a drug dose
of 0.1 pmollkg (-0.1 mg/kg), compound 8 was the most active and induced a complete tumor regression in 80% of mice compared to 50% response for compounds, 7,9, and 10. In addition, no apparent systemic toxicity was observed with compound 8 at doses as high as 10pmol/kg.
Discussion The preparation of all four SiNc 7-10 bearing different axial substituents followed straightforward synthetic procedures to yield single isomeric products. The latter is a major advantage over synthetic routes involving substitutions of various positions on the macrocycle which often results in difficult to resolve, complex isomeric mixtures. The final products were purified by conventional aluminum oxide column chromatography. The purity of the compounds was assessed by normal-phase HPLC and silica gel TLC, which revealed the various SiNc as single green spota. Fast atom bombardment mass spectrometry or combustion analyses were used to confirm the molecular composition. The SiNC derivatives 9 and 10gave adequate combustion analyses and definite melting points
